(p.154) Appendix Methodology
(p.154) Appendix Methodology
Accountants' Truth is based on in-depth interviews with employees of the largest accounting firms in London. Interviewees were asked to explain how they would account for a fictional scenario, concerning which the accounting rules were indeterminate, and then to describe similarly contentious accounting problems from their own experience. The interviews focused on the ways in which interviewees constructed accounting knowledge, but also included broader discussion of professionalism and ethics. The interviews were recorded, fully transcribed, manually coded, and interpreted qualitatively using techniques of discourse, narrative, and textual analysis.
This appendix is addressed primarily to fellow researchers who want to understand the methodological choices I made. It will discuss my selection of interviewees, the questions I asked them, the manner in which I did so, my interpretation of the results, and the limitations of the research as well as the steps I took to overcome these. It will also discuss the opportunities and challenges posed by my having shared my interviewees' professional experience. The overall objective of the research was not to understand accountants' life experience, their behaviour, or their social environment, but rather to understand the workings of an increasingly rationalized, globalized, and powerful discourse. All these are inextricably connected, but nonetheless my primary focus on accounting discourse affects all aspects of the methodology.
I conducted twenty interviews between June 2003 and October 2004. Interviewees were fully qualified chartered accountants,102 aged 35 or below, male, and employed by one of the ‘big four’ firms in London (Deloitte, Ernst & Young, KPMG, and PricewaterhouseCoopers). Five interviewees were selected from each firm. I used the snowball technique to identify them, beginning with personal acquaintances and asking each interviewee for a maximum of two follow-on contacts.103
My sampling decisions regarding gender, age, employer, and location were always in favour of the majority.104 In the United Kingdom, there are more male than female chartered accountants; the big four firms are the largest employers of them (and employ more people under 35 than over); London has a higher concentration of chartered accountants than any other city; and the Institute of Chartered Accountants in England and Wales (ICAEW) is the largest of the various institutes serving the profession.105 (p.155) Accountants meeting all these criteria at once are, nonetheless, a minority group. Yet they remain of particular interest as an archetype of the profession, embodying in concentrated form the predominant characteristics of the profession's members. These selection criteria for interviewees, together with their relatively recent socialization through the profession's three-year examination and training period,106 maximized the interviews' salience as information-rich examples of a standardized discourse.107
Relative to the field of accounting and finance taken as a whole, fully qualified chartered accountants working in the largest London firms occupy a somewhat elite position. None of my interviewees were employed to prepare accounts, or to record a company's transactions; they worked, for instance, as auditors, financial consultants, or insolvency practitioners, engaged in work that is inherently judgemental and high level. Although my interviewees were not partners overseeing teams doing the judgemental, high-level work I describe, neither were they the most junior members of those teams. Partners and other senior people, in any case, rely on other members of their teams to formulate the information on which they base their decisions. The more senior they are, the more decisions they have to make, and the greater the level of trust they must place in their subordinates. Consequential judgements are delegated to accountants like those I interviewed, in the form of dilemmas such as the following: Should I mention this to the partner? What information should I show him or her? How should I present that information, knowing that the way I present it will affect the decision he or she makes? Judgemental accounting decisions are therefore team efforts to which accountants like those I interviewed make important contributions.
More broadly, it is a sociological mistake to study only the senior people in large organizations. Senior partners, like chief executives of large companies, have limited control over their firms. This is firstly because absolute power does not pass through them but is instead diffused in complex ways through their organizations, secondly because they are themselves products of their institutional and professional environment to a large extent, and thirdly because they are constrained by social expectations of how they should perform their role (most specifically the expectations of clients and colleagues). Focusing on middle-ranking accountants (neither very junior nor very senior) enabled me to study the detail of how accounting work is done, rather than talking to people who are either too inexperienced and under-socialized to be representative, or so senior that their role is more to manage the work than to actually do it. I was therefore able to explore the accounting discourse in use, and it is the power of that discourse, not of particular individual accountants, that is significant here.
The questions I asked my interviewees were based on the interview schedule below, which I developed during my earliest interviews. However, it was no more than a prompt: I adapted my interview strategy to each encounter, using the schedule simply to ensure that I kept all the discussions relevant.
The second section of the interview schedule refers to a broad-ranging discussion that I had with each interviewee about a fictional scenario I gave to them in advance. The scenario is discussed in detail in Chapter 3. It provided a common point of comparison between interviews, and a way of focusing interviewees on the types of situation I wanted to explore. After discussing the scenario, I asked for similar situations my interviewees had been in themselves, which we discussed by way of transition to their own experience.
I asked about technical disputes my interviewees had experienced in order to explore where the crystallization of accounting facts either failed or came under pressure. Asking about intense disputes brought out the emotive and argumentative aspects of knowledge construction, and broadened the discussion beyond areas that interviewees defined as accounting matters. I focused particularly on the difficulties of communication surrounding such disputes, in order to explore the limitations of the discourse with which my interviewees were working. In the final portion of the interviews I explored what communication and negotiation of ethical positions was possible between accountants in a professional context, and what kind of ethical vocabulary they used. Unless a specific opportunity arose to discuss them earlier on, I waited until late in the interviews, when interviewees were most comfortable, before broaching these increasingly taboo and hazardous topics.
(p.157) Interview technique
Whilst interviewing, I did not try to elicit the attitudes of interviewees, but either to engage them in technical or professional disputes, or to explore practical situations they remembered. Variations in attitudes expressed in response to different lines of questioning in the same interview, as often occur, undermine attempts to identify what interviewees really think. I therefore elicited responses from interviewees that I could analyse in terms of practices and conditions of possibility rather than attitudes (Potter and Wetherell 1987).
Although I did not seek my interviewees' attitudes in the abstract, however, I did explore their motivational thinking with respect to specific actions, often by asking why. There was a danger in doing so of asking closed questions which drew out particular kinds of reasons and focused on particular actions as needing to be justified. Whilst being sensitive to this, I nonetheless chose to treat my interviewees as peers who might change my mind, rather than as specimens demanding impartial representation. ‘The interpreter does not dispense with his prejudices. He puts them at risk’ (Weinsheimer 1985, 208). I was therefore actively engaged throughout each interview, offering alternative perspectives, seeking contradictions, revisiting key questions, and challenging the positions taken by interviewees. This active approach enabled me to test the limits of my interviewees' discourse and the responses that they were able to make through it. My interview practice thus embraced, rather than resisted, the impossibility of impartiality in an interview setting. I could have made other choices, but not no choice at all. An interviewer, like an actor on stage, changes the scene simply by being there. Even standing on a stage in ‘neutral’ is very different to being offstage; it is not actually neutral, but rather a choice to interact with an audience in a particular conventional way.
Each interview lasted around ninety minutes. I interviewed my respondents on an anonymous basis at the London School of Economics (i.e. away from their places of work), in order to set them at ease about speaking to me openly. Our conversations confirmed that accountants are very sensitive about the slightest impression they create within their firms, so worries about being seen in the office with a researcher or about being overheard would have restricted their responses there. This is particularly the case given the difficulty of achieving privacy in offices that are largely open plan.
Whilst I argued with interviewees in order to draw them out when we spoke, I did not use the same strategy when analysing the interview transcripts afterwards. At that stage, my priority was to understand rather than to challenge their discourse:
My analysis was, therefore, more sympathetic to my interviewees than were the interviews themselves. I interpreted the interviews qualitatively using techniques of discourse, (p.158) narrative, and textual analysis, and in particular by asking such questions as the following: What stylistic choices has the speaker made? Is there a subtext that might alter the most obvious interpretation? How has a particular kind of utterance become possible or impossible in my interviewees' social context?108 What story is being told about the speaker, as well as his subject? What role is the interviewee playing, and how? What does the interviewee want to persuade me of, and why? These questions are drawn from distinct critical traditions, including rhetoric, deconstruction, narrative analysis, and dramatic criticism. Cumulatively, asking them enabled me to gain a multidimensional perspective on what interviewees were doing in response to my interview questions. Of course, I was also sensitive to, and interested in, how my questions themselves provoked particular kinds of response.
When we try to understand a text, we do not try to transpose ourselves into the author's mind but, if one wants to use this terminology, we try to transpose ourselves into the perspective within which he has formed his views. But this simply means that we try to understand how what he is saying could be right. If we want to understand, we will try to make his arguments even stronger. (Gadamer 1989, 292)
The question about storytelling is more helpful in interpreting accounting discourse than one might think. Although accounting facts are couched in static terms, their construction is littered with competing narratives that are suppressed in the final product. More broadly, my interviewees appealed to a range of narratives in order to persuade both themselves and others of the legitimacy and coherence of their actions and beliefs. When I asked them what they did and why, I found that they understood their current work as part of a personal narrative, with a past and a future. The importance of such narratives leads me to adopt a different emphasis to that of many previous studies. The first full-scale ethnography of accountants' training in England, for instance, adopted a symbolic interactionist approach, focusing on the day-to-day aspects of accounting life. Insofar as accountants' personal narratives are considered under this approach, they are considered as processes of socialization, whereby accountants are gradually moulded by their day-to-day practice and by the institutions relevant to their work (Coffey 1993, 44–5). The only other full-scale study of accountants' training in England to date distanced itself from symbolic interactionism to some extent, although it retained a more general phenomenological focus on situational specifics (Anderson-Gough, Grey and Robson 1998a, 36–42, 126).109 Accountants' Truth differs in its focus, taking a greater interest in questions of character, ethics, and identity, for instance. As Robert White puts it, describing the work of Henry Murray:
White, like several post-war contemporaries including Erich Fromm and David Riesman, used psychoanalytic techniques to gain insight into social trends (Fromm 1956; Riesman, Denney, and Glazer 1950). This research agenda suffered a setback, however, with the demise of psychoanalysis itself in mainstream psychology,110 and was displaced by movements such as symbolic interactionism, ethnomethodology, and phenomenology, which were more interested in social practices and institutions (Blumer 1969; Garfinkel 1967; Schutz 1967). However a concern with the relation between character and culture did not only survive in the tradition of life history interviewing (on which see Plummer 2001). In economic sociology, it persisted in the interview technique of writers such as Sennett (1998), Sennett and Cobb (1977), Margolis (1979), Hochschild (1983), and Wajcman (1983; 1998).
If small parts and short segments of human affairs have to be isolated for detailed scrutiny, they must still be understood as parts of a patterned organic system and as segments of a lifelong process. This has never meant [for Murray] that all research should take the form of collecting life histories.…It implies simply that isolating, fragmenting, and learning just a tiny bit about a lot of people tends to carry us away from what is most worth studying. (White 1963, xiii)
(p.159) My focus on accounting discourse rather than on accountants themselves, and on the accounting field rather than on society as a whole, does tend in a phenomenological direction. Although I knew some of my interviewees, most of the interviews were one-off encounters between strangers. My primary interest being in discourse rather than in individuals led me not to ask for my interviewees' overall life stories, or to seek details of their social backgrounds systematically. Nonetheless, I placed strong emphasis on White's observation that the part cannot be understood independently of the whole, exploring my interviewees' narratives and motivations as necessary to flesh out the contexts and back stories of the situations we discussed. My interest in character and ethics, and ultimately in the potential of professional identity in the accounting field, is commensurate with this difference in emphasis.
Advantages and limitations to the methodology
The general advantages and limitations of research based on in-depth interviews are well documented, and are relevant here (e.g. see Creswell 1998; King, Keohane, and Verba 1994; Lather 1991; Plummer 2001; Silverman 2001). However, there are particular reasons why in-depth interviewing suited this study. Firstly, my interest in accounting discourse made it appropriate to analyse a small number of interviews in detail, rather than using a larger sample or engaging in participant observation that would have necessitated a less detailed approach. Interviewing was better suited to studying accounting discourse than it would have been to studying how that discourse becomes established in particular institutional contexts, for instance, for which participant observation may have been preferable (e.g. see Born 2005, 97–128). Secondly, my prior experience of the field meant that I already had a lot of the background knowledge I might otherwise have gained through participant observation. Thirdly, the interviews to some extent replicated my interviewees' experience of debating accounting treatments with clients and colleagues. Therefore the artificiality of the interview context was not, in this case, as much of a drawback as it might at first seem, because the creation of accounting facts is itself an artificial process. Luke, for instance, even thanked me for my time after his interview, and was then embarrassed at having interacted with me as if I were a client.
Nonetheless, my choice of method assumes that accountants' discourse is highly standardized. If that were not the case, not only would a small sample be insufficient to draw worthwhile conclusions, but studying accounting discourse would itself not make sense. I have argued for this assumption in Chapters 1 and 4, but it was in any case supported by my experience of the field. Twenty interviews were sufficient to identify regularities in interviewees' verbal reasoning, and incremental interviews were becoming repetitive and yielding few fresh insights. Given that accountants are rigorously socialized, and that they must discursively represent commercial information in accordance with accounting standards that are scrupulous in establishing terminology, a lack of such regularities would have been surprising. Accounting language's continuity across organizations is, consequently, more interesting to me than its uniqueness in any particular firm, so I chose not to work within a case study organization. I did not restrict my sample to those engaged in a particular area of work within the big four firms, such as audit or tax, for similar reasons.111
(p.160) I do not, then, seek to prove the existence of a standardized accounting discourse through my empirical research, but instead I establish that as a well-founded assumption at the outset. My research itself is directed towards achieving and communicating insights into that discourse. I justify some further generalizations by reference to the constraints my interviewees worked within, as with my discussion of what it was possible for them to say, and what roles they could adopt. Beyond this, however, it is only possible to make a persuasive but not conclusive argument for the broader relevance of research such as this. Patti Lather articulates some of the strategies that I have used to do so: I have sought to demonstrate ‘triangulation validity’ by reference to a range of relevant literatures and to written accounting standards and professional guidance, as well as to my interview material; ‘construct validity’ by looking for what counts as fact within accountancy rather than imposing external criteria; ‘face validity’ by making my interpretation of a specialist area as readily accessible to non-specialists as possible; and ‘catalytic validity’ by openly conveying my normative position with respect to the research and arguing for its importance on that basis (Lather 1991, 66–9). I am of course not solely describing accountancy but, inevitably, also participating in its ongoing social construction. The accounting discourse itself does not offer an impartial impression of social reality in any case, but is worthy of analysis precisely because it does not.
Readers familiar with work in other professional service environments, particularly consulting and law, will see many important parallels with accounting. Unfortunately, it is beyond the scope of this book to make detailed comparisons to other fields, although I do situate accounting within the broader social context it shares with them. My argument here relies neither on generalizability beyond nor uniqueness to accounting. It simply claims validity with respect to accounting, which matters as a specific case because of its central role in economic life.
Interview research as an insider
I conducted and interpreted the interviews as an insider, having spent four years training and working as a chartered accountant with PricewaterhouseCoopers in London between 1998 and 2002. The characteristics and experience according to which I selected my interviewees are characteristics and experience that I share. My interviewees were therefore not only my peers but, in some schematic respects, they were those of my peers who were most like me. I found that accountants were more at ease talking to me in the knowledge that I shared their professional background, so I made no attempt to hide that fact. Our shared experience was, in any case, our basis for discussion: it would have been difficult for a non-accountant to gain access to my interviewees, or to debate accounting matters with them in the detail necessary for this study.
I argued in the previous section that my personal experience of accountancy compensates in some ways for the absence of an ethnographic element to the research, but in reality it does both less and more than this. It does less in that I was not consciously researching during my time in the field, and so did not record my experiences or analyse them sociologically. I was not attempting to maintain objectivity, although I have argued above that this is a questionable ideal. Because most of my interviews were conducted with people who had not previously been my colleagues, I did not develop the trust and rapport with my interviewees that ethnographers gain. My experience gave (p.161) me more insight than an ethnography could have done, however, in that I lived it for real, rather than as a temporary foray engaged in for research purposes. As both my own research and the other research I refer to during the book attest, the experience of young accountants in the large firms is an intense and demanding one. The discipline, competitiveness, insecurity, and camaraderie it breeds are difficult to access without having experienced the stakes involved when, for instance, your first job and career depend on passing a competitive exam, or your sense of self-worth becomes bound up with your colleagues' appraisal of your interpersonal impact. This is a visceral experience that cannot be readily observed from without. Having shared it with my interviewees gave me a heightened sense for which questions mattered most as our conversations developed, and for what their answers meant.
It might be asked why, having shared my interviewees' professional experience, I would want to ask them about it in interviews at all. Surely I already knew the answers to my questions? This objection is a challenge to anyone conducting research as an insider. Yet the kind of knowledge I shared with my interviewees was an inarticulate practical knowledge. The interviews created a kind of dialogue which does not ordinarily take place, and which made that knowledge explicit and debatable. The interviews were important not for what my interviewees could tell me that I did not already know, therefore, but for their witting or unwitting collaboration in an attempt, through dialogue and argument, to find ways to articulate, debate, and criticize our shared experience and practice.
(102.) Most of my interviewees qualified with the ICAEW, so my discussion focuses on that institute. However Ernst & Young encouraged its trainees to become members of ICAS instead of the ICAEW in some years, so it was impractical to distinguish between ICAEW- and ICAS-trained employees of that firm. One interviewee (Jason) had trained with CIPFA but had since moved to the private sector.
(103.) Many of those I asked did not agree to be interviewed, so the sample was self-selecting to some extent. My emphasis on the common aspects of interviewees' shared discourse, however, minimizes this limitation.
(104.) I could have extended my sampling restrictions to cover race (English only) and class (middle class or aspiring middle class only). However, the dividing lines here are insufficiently clear to do so with any confidence, so I chose not to claim selection criteria which I could not uphold in practice.
(105.) Although 45 per cent of ICAEW trainees were female in 2002, only 20 per cent of qualified members were female (Financial Reporting Council 2003, 32, 43). The big four accounted for 46 per cent of all ICAEW trainees at 31 July 2003, and the figure would have been higher but for the fact that Ernst & Young had recently used ICAS as a substitute for the ICAEW in training many of their staff (ICAEW 2003). The fee income of the smallest of the big four in 2002/03, Ernst & Young (£755 million), was three-and-a-half times that of its nearest rival, Grant Thornton (£216 million) (Fisher 2003, 27). In 2002/03, 40 per cent of new UK students joined the London district society of the ICAEW, as compared to 5 per cent joining the Birmingham and West Midlands society (Birmingham being the United Kingdom's second largest city) (ICAEW 2003). The predominance of London remains but is less marked for ICAEW members as a whole, of whom 26 per cent were attached to the London district society on 1 April 2004, suggesting that London accountants are younger than average (ICAEW 2004). Evidence for the youth of ICAEW trainees in general, and big four trainees in particular, has already been cited in n. 9. In the United Kingdom and Ireland, the ICAEW had more than double the number of members of ACCA, its nearest rival, in 2002 (Financial Reporting Council 2003, 26).
(106.) For evidence of the striking effectiveness of this socialization, see the existing ethnographic studies of chartered accountancy training in the United Kingdom (Anderson-Gough, Grey, and Robson 1998 a; Coffey 1993; Power 1997).
(107.) Of course this is not to endorse the age, gender, and regional biases in British accounting, but merely to recognize their potential salience with respect to the profession's dominant discourse. Further research would be needed to determine their actual salience.
(108.) The question of possibility in discourse is Michel Foucault's. He suggests that ‘[e]ach society has its regime of truth, its “general politics” of truth: that is, the types of discourse which it accepts and makes function as true;…the status of those who are charged with saying what counts as true’ (1980 a, 131). To ask about conditions of possibility, in this study, is to ask both how interviewees are enabled and how they are restricted by the politics of truth in accounting. What resources (p.173) are available to them to make themselves believed? What would not be believed if they said it? How have they become able, and unable, to think?
(109.) Although these studies of accountants' training are important examples, the picture is more complex when general ethnographies of accountancy practices, or interview-based studies, are considered. Harper (1988), for instance, adopts Goffman's interest in performance, and Empson (2004) focuses on the relation between organizational and professional identity.
(110.) For a recent discussion of the place of psychoanalysis in various disciplines, see Brooks and Woloch's Whose Freud? (2000). Psychoanalysis remains a popular paradigm in literary studies, where its emphasis on life histories can be translated onto the discussion of fictional lives and of narrative more generally (Brooks 1994; Mitchell 1981). The influence of the psychoanalytic tradition on my own interpretative practice, as described above, is clearly mediated through narrative analysis.
(111.) Most interviewees trained as auditors, but many had moved to different areas of their firms by the time I interviewed them. Anderson-Gough, Grey, and Robson found during their ethnography that although ‘socialization processes vary between divisions’, the claim that these processes ‘give rise to differing conceptions of the role of the professional accountant’ is ‘largely false’ (1998 a, 5, 135–6).